Vision Arena

View rankings across multimodal, generative AI models capable of understanding and processing visual inputs.

Last Updated

Oct 6, 2025

Total Votes

532,681

Total Models

/


1	gemini-2.5-pro	1241	±8	62,824	Google	Proprietary
1	chatgpt-4o-latest-20250326	1234	±8	14,069	OpenAI	Proprietary
2	gpt-4.5-preview-2025-02-27	1220	±11	2,925	OpenAI	Proprietary
3	o3-2025-04-16	1213	±8	42,177	OpenAI	Proprietary
3	gpt-5-chat	1212	±9	35,271	OpenAI	Proprietary
3	gpt-4.1-2025-04-14	1207	±8	38,606	OpenAI	Proprietary
3	gemini-2.5-flash	1203	±8	37,468	Google	Proprietary
3	claude-opus-4-20250514-thinking-16k	1203	±15	1,430	Anthropic	Proprietary
3	gemini-2.5-flash-preview-09-2025	1201	±16	1,431	Google	Proprietary
3	hunyuan-vision-1.5-thinking	1200	±14	1,954	Tencent	Proprietary
3	claude-sonnet-4-20250514-thinking-32k	1198	±16	1,321	Anthropic	Proprietary
3	qwen3-vl-235b-a22b-instruct	1197	±15	1,845	Alibaba	Apache 2.0
5	gpt-5-high	1194	±9	29,364	OpenAI	Proprietary
5	qwen3-vl-235b-a22b-thinking	1189	±15	1,657	Alibaba	Apache 2.0
5	claude-3-7-sonnet-20250219-thinking-32k	1189	±15	1,649	Anthropic	Proprietary
6	o4-mini-2025-04-16	1194	±8	38,267	OpenAI	Proprietary
6	gpt-4.1-mini-2025-04-14	1194	±8	39,587	OpenAI	Proprietary
7	o1-2024-12-17	1187	±10	3,694	OpenAI	Proprietary
8	claude-opus-4-20250514	1183	±12	2,536	Anthropic	Proprietary
8	claude-sonnet-4-20250514	1181	±14	2,052	Anthropic	Proprietary
8	grok-4-0709	1179	±9	28,500	xAI	Proprietary
13	qwen-vl-max-2025-08-13	1172	±13	3,058	Alibaba	Proprietary
14	gpt-5-mini-high	1173	±10	27,181	OpenAI	Proprietary
16	Gemini-1.5-Pro-002	1172	±8	8,916	Google	Proprietary
23	gpt-4o-2024-05-13	1156	±8	23,280	OpenAI	Proprietary
23	glm-4.5v	1151	±13	3,168	Z.ai	MIT
24	Claude 3.5 Sonnet (10/22)	1155	±7	10,549	Anthropic	Proprietary
24	step-1o-turbo-202506	1149	±14	2,069	StepFun	Proprietary
25	mistral-medium-2505	1149	±9	11,258	Mistral	Proprietary
25	gemma-3-27b-it	1146	±9	16,737	Google	Gemma
25	Claude 3.5 Sonnet (06/20)	1141	±9	21,632	Anthropic	Proprietary
25	hunyuan-large-vision	1140	±16	1,470	Tencent	Proprietary
25	gpt-5-nano-high	1139	±12	3,844	OpenAI	Proprietary
25	step-3	1137	±13	3,169	StepFun	Apache 2.0
27	mistral-medium-2508	1137	±9	30,466	Mistral	Proprietary
27	llama-4-maverick-17b-128e-instruct	1136	±10	6,901	Meta	Llama 4
27	Gemini-1.5-Flash-002	1132	±9	7,256	Google	Proprietary
27	mistral-small-2506	1131	±10	11,521	Mistral	Apache 2.0
27	claude-3-5-haiku-20241022	1128	±15	1,579	Anthropic	Propretary
29	gemini-2.0-flash-lite-preview-02-05	1127	±10	4,001	Google	Proprietary
32	llama-4-scout-17b-16e-instruct	1119	±10	6,393	Meta	Llama
32	step-1o-vision-32k-highres	1118	±12	2,844	StepFun	Proprietary
34	mistral-small-3.1-24b-instruct-2503	1117	±9	27,486	Mistral	Apache 2.0
34	qwen2.5-vl-72b-instruct	1116	±10	3,767	Alibaba	Qwen
35	GPT-4o (08/06)	1113	±12	3,384	OpenAI	Proprietary
35	qwen2.5-vl-32b-instruct	1110	±16	1,490	Alibaba	Apache 2.0
38	gemini-1.5-pro-001	1112	±11	16,735	Google	Proprietary
39	gpt-4-turbo-2024-04-09	1107	±11	13,390	OpenAI	Proprietary
46	gpt-4.1-nano-2025-04-14	1083	±18	1,211	OpenAI	Proprietary
47	GPT-4o-mini (07/18)	1092	±7	17,362	OpenAI	Proprietary
48	pixtral-large-2411	1088	±8	5,439	Mistral	MRL
48	qwen-vl-max-1119	1079	±15	1,431	Alibaba	Proprietary
49	qwen2-vl-72b	1080	±9	5,946	Alibaba	Qwen
51	gemini-1.5-flash-8b-001	1064	±10	6,255	Google	Proprietary
51	step-1v-32k	1056	±16	1,551	StepFun	Proprietary
52	claude-3-opus-20240229	1059	±9	15,568	Anthropic	Proprietary
52	gemini-1.5-flash-001	1054	±11	13,261	Google	Proprietary
54	hunyuan-standard-vision-2024-12-31	1038	±21	809	Tencent	Proprietary
55	molmo-72b-0924	1042	±12	3,063	AI2	Apache 2.0
58	llama-3.2-vision-90b-instruct	1027	±8	8,708	Meta	Llama 3.2
58	qwen2-vl-7b-instruct	1026	±10	5,778	Aliaba	Apache 2.0
58	internvl2-26b	1019	±12	5,146	OpenGVLab	MIT
59	pixtral-12b-2409	1019	±9	7,520	Mistral	Apache 2.0
59	amazon-nova-lite-v1.0	1015	±14	1,856	Amazon	Proprietary
59	amazon-nova-pro-v1.0	1015	±13	2,338	Amazon	Proprietary
59	claude-3-sonnet-20240229	1014	±11	12,314	Anthropic	Proprietary
61	c4ai-aya-vision-32b	995	±22	847	Cohere	CC-BY-NC-4.0
62	yi-vision	999	±18	1,220	01 AI	Proprietary
63	claude-3-haiku-20240307	997	±12	13,380	Anthropic	Proprietary
64	molmo-7b-d-0924	991	±13	2,828	AI2	Apache 2.0
65	nvila-internal-15b-v1	982	±20	1,078	NVIDIA	-
67	llama-3.2-vision-11b-instruct	985	±10	4,818	Meta	Llama 3.2
67	llava-onevision-qwen2-72b-ov	975	±18	1,321	LLaVA	Apache 2.0
70	minicpm-v-2_6	959	±15	1,987	OpenBMB	Apache 2.0
71	cogvlm2-llama3-chat-19b	958	±15	1,991	Zhipu AI	CogVLM2
72	llava-v1.6-34b	961	±12	4,531	LLaVA	Apache 2.0
72	internvl2-4b	952	±12	3,706	OpenGVLab	MIT
78	phi-3.5-vision-instruct	917	±15	2,592	Microsoft	MIT
79	phi-3-vision-128k-instruct	878	±18	1,401	Microsoft	MIT

Take your chats anywhere

Vision Arena

Default Leaderboard Plots

Battle Count for Each Combination of Models (without Ties)

Average Win Rate Against All Other Models (Uniform Sampling and No Ties)

Confidence Intervals on Model Strength (via Bootstrapping)

Fraction of Model A Wins for All Non-tied A vs. B Battles

Vision Arena

Default Leaderboard Plots

Battle Count for Each Combination of Models (without Ties)

Average Win Rate Against All Other Models (Uniform Sampling and No Ties)

Confidence Intervals on Model Strength (via Bootstrapping)

Fraction of Model A Wins for All Non-tied A vs. B Battles